Enter opens the highlighted result. Esc closes.

Nemotron 3 Ultra vs Qwen3 235B A22B Thinking 2507

Nemotron 3 Ultra leads Qwen3 235B A22B Thinking 2507 on the Artificial Analysis Intelligence Index, 22.9 to 12.7, and Nemotron 3 Ultra costs 1.5x less per output token ($2.40 vs $3.50 per million). Updated October 23, 2026; 41 benchmarks have a result for both.

Replace Nemotron 3 Ultra Replace Qwen3 235B A22B Thinking 2507
Nemotron 3 UltraNVIDIA
Coding
30.0
Overall
53.1
Output $/M
$2.40
38
Qwen3 235B A22B Thinking 2507Qwen
Coding
No data
Overall
52.9
Output $/M
$3.50
2
vs boards won of 41 shared, 1 tied

Nemotron 3 Ultra or Qwen3 235B A22B Thinking 2507: which to pick

Only Nemotron 3 Ultra has a coding score, 30.0 on the Vibe Coding Index; Qwen3 235B A22B Thinking 2507 has none, so the two cannot be ranked on coding.

Pick Qwen3 235B A22B Thinking 2507 for work measured by LMArena Text: medicine and healthcare or LMArena Text: Polish, where it leads.

Model ANemotron 3 UltraNVIDIA30.0/1003 independent benchmarksNo cost per solved taskModel BQwen3 235B A22B Thinking 2507QwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
41
Nemotron 3 Ultra only
30
Qwen3 235B A22B Thinking 2507 only
5
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Nemotron 3 Ultra 30.0 from 3 benchmarks, Qwen3 235B A22B Thinking 2507 not benchmarked.
Widest category
Web design, 18.9 points apart on average over 6 shared benchmarks.
Widest single gap
AA Coding Index: Nemotron 3 Ultra 49.3 against Qwen3 235B A22B Thinking 2507 22.1, 27.2 points apart on a 0 to 100 scale.
Output price
Nemotron 3 Ultra $2.40 against Qwen3 235B A22B Thinking 2507 $3.50 per million output tokens, 1.5x apart.
Context window
Nemotron 3 Ultra takes 262K tokens, Qwen3 235B A22B Thinking 2507 128K.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryNemotron 3 UltraBarsQwen3 235B A22B Thinking 2507Ties
Coding40
Intelligence10
Agentic work10
Reasoning50
Math201 tie
Knowledge31
Web design60
General161
All shared boards3821 tie
CodingLike-for-likeSix of these boards feed the Vibe Coding Index40
Coding benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
AA Coding Index49.322.1Nemotron 3 Ultra
LiveCodeBench (Vals AI)86.0%Not measuredOnly Nemotron 3 Ultra measured
SWE-bench Verified (Vals AI)69.0%Not measuredOnly Nemotron 3 Ultra measured
Vibe Code Bench (Vals AI)7.6%Not measuredOnly Nemotron 3 Ultra measured
Code migration (Vals AI)4.9%Not measuredOnly Nemotron 3 Ultra measured
ProgramBench (Vals AI)0.0%Not measuredOnly Nemotron 3 Ultra measured
Terminal-Bench 2.1 (Vals AI)50.9%Not measuredOnly Nemotron 3 Ultra measured
SciCode40.3%Not measuredOnly Nemotron 3 Ultra measured
Terminal-Bench Hard (Artificial Analysis run)36.4%Not measuredOnly Nemotron 3 Ultra measured
FrontierCode13.6%Not measuredOnly Nemotron 3 Ultra measured
WeirdML43.5%41.0%Nemotron 3 Ultra
LMArena Text: coding14671424Nemotron 3 Ultra
LMArena Text: software and IT14681431Nemotron 3 Ultra
Neither model has a result on the other 30 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
AA Intelligence Index22.912.7Nemotron 3 Ultra
Epoch Capabilities Index146.0Not measuredOnly Nemotron 3 Ultra measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
Agentic work benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
AA Agentic Index20.11.3Nemotron 3 Ultra
Finance Agent (Vals AI)37.7%Not measuredOnly Nemotron 3 Ultra measured
Legal Research Bench (Vals AI)15.4%Not measuredOnly Nemotron 3 Ultra measured
Excel Modeling Benchmark (Vals AI)31.7%Not measuredOnly Nemotron 3 Ultra measured
τ²-Bench Telecom83.3%Not measuredOnly Nemotron 3 Ultra measured
APEX-Agents22.7%Not measuredOnly Nemotron 3 Ultra measured
Vending-Bench 2Not measured-$11Only Qwen3 235B A22B Thinking 2507 measured
MCP Atlas63.1%Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 34 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index50
Reasoning benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
GPQA Diamond86.1%Not measuredOnly Nemotron 3 Ultra measured
GPQA Diamond (Artificial Analysis run)86.7%Not measuredOnly Nemotron 3 Ultra measured
Humanity's Last Exam (Artificial Analysis run)28.4%Not measuredOnly Nemotron 3 Ultra measured
GPQA Diamond (Epoch AI run)85.4%80.1%Nemotron 3 Ultra
Mystery game puzzles20.0%9.0%Nemotron 3 Ultra
LMArena Text: hard prompts14521416Nemotron 3 Ultra
LMArena Text: expert14721463Nemotron 3 Ultra
LMArena Text: hard prompts in English14531432Nemotron 3 Ultra
Neither model has a result on the other 15 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
OTIS Mock AIME86.7%86.7%Tie
ProofBench2.0%Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: math14571412Nemotron 3 Ultra
LMArena Text: mathematical work14621435Nemotron 3 Ultra
Neither model has a result on the other 10 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index31
Knowledge benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
MMLU Pro85.8%Not measuredOnly Nemotron 3 Ultra measured
LegalBench (Vals AI)82.1%Not measuredOnly Nemotron 3 Ultra measured
CorpFin (Vals AI)65.5%Not measuredOnly Nemotron 3 Ultra measured
TaxEval (Vals AI)73.1%Not measuredOnly Nemotron 3 Ultra measured
MedCode (Vals AI)38.6%Not measuredOnly Nemotron 3 Ultra measured
SimpleQA VerifiedNot measured40.4%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: science14651433Nemotron 3 Ultra
LMArena Text: medicine and healthcare14551456Qwen3 235B A22B Thinking 2507
LMArena Text: legal and government14531415Nemotron 3 Ultra
LMArena Text: business and finance14351412Nemotron 3 Ultra
Neither model has a result on the other 10 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index60
Web design benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
Design Arena: websites11451052Nemotron 3 Ultra
Design Arena: UI components1135942Nemotron 3 Ultra
Design Arena: game dev1138956Nemotron 3 Ultra
Design Arena: data viz1142950Nemotron 3 Ultra
Design Arena: 3D1132998Nemotron 3 Ultra
Design Arena: all code categories11461032Nemotron 3 Ultra
Design Arena: SVG1069Not measuredOnly Nemotron 3 Ultra measured
Design Arena: ASCII art1100Not measuredOnly Nemotron 3 Ultra measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index161
General benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
IFBench81.4%Not measuredOnly Nemotron 3 Ultra measured
Long context reasoning (AA-LCR)79.3%Not measuredOnly Nemotron 3 Ultra measured
DTBench90.1%80.3%Nemotron 3 Ultra
LMCA36.9%29.3%Nemotron 3 Ultra
Fiction.LiveBenchNot measured75.0%Only Qwen3 235B A22B Thinking 2507 measured
Creative writing (Lech Mazur)Not measured82.4%Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text14451415Nemotron 3 Ultra
LMArena Text: instruction following14181386Nemotron 3 Ultra
LMArena Text: creative writing13981384Nemotron 3 Ultra
LMArena Text: English14541429Nemotron 3 Ultra
LMArena Text: non-English14271397Nemotron 3 Ultra
LMArena Text: Chinese14981481Nemotron 3 Ultra
LMArena Text: German14741408Nemotron 3 Ultra
LMArena Text: JapaneseNot measured1386Only Qwen3 235B A22B Thinking 2507 measured
LMArena Text: Korean13851375Nemotron 3 Ultra
LMArena Text: French1473Not measuredOnly Nemotron 3 Ultra measured
LMArena Text: Spanish14541389Nemotron 3 Ultra
LMArena Text: Russian14171400Nemotron 3 Ultra
LMArena Text: Polish13851390Qwen3 235B A22B Thinking 2507
LMArena Text: multi-turn14111406Nemotron 3 Ultra
LMArena Text: longer queries14371399Nemotron 3 Ultra
LMArena Text: writing and language14051388Nemotron 3 Ultra
LMArena Text: entertainment and media13971370Nemotron 3 Ultra
Neither model has a result on the other 25 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
BenchmarkNemotron 3 UltraQwen3 235B A22B Thinking 2507Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Nemotron 3 Ultra$0.0018Fits in one request
Qwen3 235B A22B Thinking 2507$0.0022Fits in one request

Nemotron 3 Ultra costs less

Repository review

50K input + 3K output tokens

Nemotron 3 Ultra$0.037Fits in one request
Qwen3 235B A22B Thinking 2507$0.033Fits in one request

Qwen3 235B A22B Thinking 2507 costs less

Agent loop

200K input + 10K output tokens

Nemotron 3 Ultra$0.144Fits in one request
Qwen3 235B A22B Thinking 2507$0.125Exceeds the context window

Qwen3 235B A22B Thinking 2507 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507
SpecificationNVIDIA Nemotron 3 UltraQwen3 235B A22B Thinking 2507
MakerNVIDIAQwen
WeightsOpen weightsOpen weights
ReleasedJune 4, 2026July 25, 2025
Context window262K128K
Input $/Mtok$0.60$0.45
Output $/Mtok$2.40$3.50
Coding index30.0Not benchmarked
Independent measurements behind it30
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Nemotron 3 Ultra better than Qwen3 235B A22B Thinking 2507 for coding?

Only Nemotron 3 Ultra has a coding score, 30.0 on the Vibe Coding Index; Qwen3 235B A22B Thinking 2507 has none, so the two cannot be ranked on coding. Pick Qwen3 235B A22B Thinking 2507 for work measured by LMArena Text: medicine and healthcare or LMArena Text: Polish, where it leads.

Which is cheaper, Nemotron 3 Ultra or Qwen3 235B A22B Thinking 2507?

Nemotron 3 Ultra costs $0.60 per million input tokens and $2.40 per million output tokens; Qwen3 235B A22B Thinking 2507 costs $0.45 and $3.50. Cost per solved task is known for neither of them.

Which has the bigger context window, Nemotron 3 Ultra or Qwen3 235B A22B Thinking 2507?

Nemotron 3 Ultra, with 262K tokens against 128K.

How many benchmarks is the Nemotron 3 Ultra vs Qwen3 235B A22B Thinking 2507 comparison based on?

41 benchmarks have a result for both models as of October 23, 2026. Nemotron 3 Ultra's coding index rests on 3 independent benchmarks and Qwen3 235B A22B Thinking 2507's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next